


**********************Figure 1. Trends in decentralization over years******************


gen political= n_representation+ n_lawmaking+ n_constitutional
gen administrative= n_instdepth+ n_policyautonomy+ n_execcontrol
gen fiscal= n_fiscalautonomy+ n_borrowautonomy+ n_fiscalcontrol+ n_borrowcontrol

gen fisautonomy=n_fiscalautonomy+ n_borrowautonomy
gen fiscontrol=n_fiscalcontrol+ n_borrowcontrol

by year, sort : egen float RAI_yearly = mean(n_RAI)

by year, sort : egen float political_yearly = mean(political)
by year, sort : egen float administrative_yearly = mean(administrative)
by year, sort : egen float fiscal_yearly = mean(fiscal)

line RAI_yearly year if year>1960&year<2019, sort yaxis(1) ||line fiscal_yearly year if year>1960&year<2019, sort yaxis(2) || line administrative_yearly year if year>1960&year<2019, sort yaxis(2)||line political_yearly year if year>1960&year<2019, sort yaxis(2)  scheme(s2mono) graphregion(fcolor(white) ilcolor(white) lcolor(white))

*******Figure 2. Country-specific patterns of over-time change of decentralization*******

twoway (line fiscal year) if year>1960&year<2019 &scode==230
graph save spain, replace
twoway (line fiscal year) if year>1960&year<2019 &scode==380
graph save sweden, replace
twoway (line administrative year) if year>1960&year<2019 &scode==235
graph save portugal, replace
twoway (line administrative year) if year>1960&year<2019 &scode==770
graph save paks, replace

graph combine spain.gph sweden.gph  portugal.gph paks.gph, row(2) ycommon scheme(s2mono) graphregion(fcolor(white) ilcolor(white) lcolor(white))
graph save countryvariation, replace




***Figure 3. Number of natural disasters over time***

by year, sort : egen float disastersd_yearly = mean(z_rowmax_5dis_maj)
by year, sort : egen float earthquake_yearly = mean(row_max_n_earth_mag5)
by year, sort : egen float storm_yearly = mean(row_max_n_storm)
by year, sort : egen float flood_yearly = mean(row_max_n_spi_fl)
by year, sort : egen float draught_yearly = mean(row_max_n_spi_dr)
by year, sort : egen float volcano_yearly = mean(row_max_n_vol_onek)
 
 
line earthquake_yearly year if year>1960&year<2019, sort||lowess earthquake_yearly year if year>1960&year<2019 , scheme(s2mono) graphregion(fcolor(white) ilcolor(white) lcolor(white))
graph save earthquakeyearly, replace 
line storm_yearly year if year>1960&year<2019, sort ||lowess storm_yearly year if year>1960&year<2019, scheme(s2mono) graphregion(fcolor(white) ilcolor(white) lcolor(white))
graph save stormyearly, replace
line volcano_yearly year if year>1960&year<2019, sort ||lowess volcano_yearly year if year>1960&year<2019, scheme(s2mono) graphregion(fcolor(white) ilcolor(white) lcolor(white))
graph save volcanoyearly, replace
line flood_yearly year if year>1960&year<2019, sort ||lowess flood_yearly year if year>1960&year<2019, scheme(s2mono) graphregion(fcolor(white) ilcolor(white) lcolor(white))
graph save floodyearly, replace
line draught_yearly year if year>1960&year<2019, sort ||lowess draught_yearly year if year>1960&year<2019, scheme(s2mono) graphregion(fcolor(white) ilcolor(white) lcolor(white))
graph save draughtyearly, replace
line disastersd_yearly year if year>1960&year<2019, sort ||lowess disastersd_yearly year if year>1960&year<2019, scheme(s2mono) graphregion(fcolor(white) ilcolor(white) lcolor(white))
graph save disastersd, replace

graph combine earthquakeyearly.gph stormyearly.gph volcanoyearly.gph floodyearly.gph draughtyearly.gph disastersd.gph, row(2) scheme(s2mono) graphregion(fcolor(white) ilcolor(white) lcolor(white))





*******Appendix 1 summary statistics****
asdoc  sum fiscal n_fiscalautonomy n_borrowautonomy n_fiscalcontrol n_borrowcontrol  administrative n_instdepth n_policyautonomy n_execcontrol    row_max_n_earth_mag5 row_max_n_storm row_max_n_vol row_max_n_spi_dr row_max_n_spi_fl z_rowmax_5dis_maj z_rowmax_5dis_maj3 z_rowmax_5dis_maj5 z_rowmax_5dis_maj10 zmean_distance_5dis_cap c_zmean_distance_5dis_cap3 c_zmean_distance_5dis_cap5 c_zmean_distance_5dis_cap10 c_z_dispersion_5dis c_z_dispersion_5dis3 c_z_dispersion_5dis5 c_z_dispersion_5dis10  lnpop polity2  lngdppc gini_disp  EFindex lnresource pop65 if n_RAI!=. &year>1960&year<2019, replace








**set controls, a short list and a full list**********


global controls "  l.lnpop  l.lngdppc l.polity2"


global controls1  "  l.lnpop l.lngdppc l.polity2 l.pop65 l.input_gini_disp l.input_EFindex  l.lninput_resource "





*********estimating the average effect of disasters on centralization ********


local i = 1
foreach depvar in fiscal administrative  {
foreach indepvar in l.z_rowmax_5dis_maj l.z_rowmax_5dis_maj2 l.z_rowmax_5dis_maj3 l.z_rowmax_5dis_maj4 l.z_rowmax_5dis_maj5 l.z_rowmax_5dis_maj6 l.z_rowmax_5dis_maj7 l.z_rowmax_5dis_maj8 l.z_rowmax_5dis_maj9 l.z_rowmax_5dis_maj10{



reghdfe  `depvar' `indepvar' ${controls} i.scode##c.year , absorb(scode year) cluster (scode)
est store model_`i'
est save model_`i', replace
local i = `i'+1

}
}


****Figure 5. Average effect of disaster frequency on decentralization******
coefplot model_1 model_2 model_3 model_4 model_5 model_6 model_7 model_8 model_9 model_10 , keep (L.z_rowmax_5dis_maj L.z_rowmax_5dis_maj2 L.z_rowmax_5dis_maj3 L.z_rowmax_5dis_maj4 L.z_rowmax_5dis_maj5 L.z_rowmax_5dis_maj6 L.z_rowmax_5dis_maj7 L.z_rowmax_5dis_maj8 L.z_rowmax_5dis_maj9 L.z_rowmax_5dis_maj10) yline(0)  scheme(s2mono) graphregion(fcolor(white) ilcolor(white) lcolor(white)) vertical coeflabels (L.z_rowmax_5dis_maj="1" L.z_rowmax_5dis_maj2="2"  L.z_rowmax_5dis_maj3="3" L.z_rowmax_5dis_maj4="4" L.z_rowmax_5dis_maj5="5" L.z_rowmax_5dis_maj6="6" L.z_rowmax_5dis_maj7="7" L.z_rowmax_5dis_maj8="8" L.z_rowmax_5dis_maj9="9"  L.z_rowmax_5dis_maj10="10" ) title(fiscal) nokey
graph save fiscalcountappend, replace

coefplot model_11 model_12 model_13 model_14 model_15 model_16 model_17 model_18 model_19 model_20, keep (L.z_rowmax_5dis_maj L.z_rowmax_5dis_maj2 L.z_rowmax_5dis_maj3 L.z_rowmax_5dis_maj4 L.z_rowmax_5dis_maj5 L.z_rowmax_5dis_maj6 L.z_rowmax_5dis_maj7 L.z_rowmax_5dis_maj8 L.z_rowmax_5dis_maj9 L.z_rowmax_5dis_maj10) yline(0)  scheme(s2mono) graphregion(fcolor(white) ilcolor(white) lcolor(white)) vertical coeflabels (L.z_rowmax_5dis_maj="1" L.z_rowmax_5dis_maj2="2"  L.z_rowmax_5dis_maj3="3" L.z_rowmax_5dis_maj4="4" L.z_rowmax_5dis_maj5="5" L.z_rowmax_5dis_maj6="6" L.z_rowmax_5dis_maj7="7" L.z_rowmax_5dis_maj8="8" L.z_rowmax_5dis_maj9="9"  L.z_rowmax_5dis_maj10="10" ) title(administrative) nokey
graph save admincountappend, replace



graph combine fiscalcountappend.gph admincountappend.gph , row(1) ycommon scheme(s2mono) graphregion(fcolor(white) ilcolor(white) lcolor(white))


graph save counteffectappend, replace


*********appendix 3 Average effect of disaster frequency on decentralization for all moving averages with a full list of control variables*****
local i = 1
foreach depvar in fiscal administrative  {
foreach indepvar in l.z_rowmax_5dis_maj l.z_rowmax_5dis_maj2 l.z_rowmax_5dis_maj3 l.z_rowmax_5dis_maj4 l.z_rowmax_5dis_maj5 l.z_rowmax_5dis_maj6 l.z_rowmax_5dis_maj7 l.z_rowmax_5dis_maj8 l.z_rowmax_5dis_maj9 l.z_rowmax_5dis_maj10{



reghdfe  `depvar' `indepvar' ${controls1} i.scode##c.year , absorb(scode year) cluster (scode)
est store modela_`i'
est save modela_`i', replace
local i = `i'+1

}
}

coefplot modela_1 modela_2 modela_3 modela_4 modela_5 modela_6 modela_7 modela_8 modela_9 modela_10 , keep (L.z_rowmax_5dis_maj L.z_rowmax_5dis_maj2 L.z_rowmax_5dis_maj3 L.z_rowmax_5dis_maj4 L.z_rowmax_5dis_maj5 L.z_rowmax_5dis_maj6 L.z_rowmax_5dis_maj7 L.z_rowmax_5dis_maj8 L.z_rowmax_5dis_maj9 L.z_rowmax_5dis_maj10) yline(0)  scheme(s2mono) graphregion(fcolor(white) ilcolor(white) lcolor(white)) vertical coeflabels (L.z_rowmax_5dis_maj="1" L.z_rowmax_5dis_maj2="2"  L.z_rowmax_5dis_maj3="3" L.z_rowmax_5dis_maj4="4" L.z_rowmax_5dis_maj5="5" L.z_rowmax_5dis_maj6="6" L.z_rowmax_5dis_maj7="7" L.z_rowmax_5dis_maj8="8" L.z_rowmax_5dis_maj9="9"  L.z_rowmax_5dis_maj10="10" ) title(administrative) nokey title(fiscal) nokey
graph save fiscalcountappenda, replace

coefplot modela_11 modela_12 modela_13 modela_14 modela_15 modela_16 modela_17 modela_18 modela_19 modela_20, keep (L.z_rowmax_5dis_maj L.z_rowmax_5dis_maj2 L.z_rowmax_5dis_maj3 L.z_rowmax_5dis_maj4 L.z_rowmax_5dis_maj5 L.z_rowmax_5dis_maj6 L.z_rowmax_5dis_maj7 L.z_rowmax_5dis_maj8 L.z_rowmax_5dis_maj9 L.z_rowmax_5dis_maj10) yline(0)  scheme(s2mono) graphregion(fcolor(white) ilcolor(white) lcolor(white)) vertical coeflabels (L.z_rowmax_5dis_maj="1" L.z_rowmax_5dis_maj2="2"  L.z_rowmax_5dis_maj3="3" L.z_rowmax_5dis_maj4="4" L.z_rowmax_5dis_maj5="5" L.z_rowmax_5dis_maj6="6" L.z_rowmax_5dis_maj7="7" L.z_rowmax_5dis_maj8="8" L.z_rowmax_5dis_maj9="9"  L.z_rowmax_5dis_maj10="10" ) title(administrative) nokey title(administrative) nokey
graph save admincountappenda, replace



graph combine fiscalcountappenda.gph admincountappenda.gph , row(1) ycommon scheme(s2mono) graphregion(fcolor(white) ilcolor(white) lcolor(white))


graph save counteffectappenda, replace


*******Appendix 4. Estimation of the effect of disaster frequency on decentralization of a selective set of models*******

****Appendix 4-1: Results of models with a short list of control variables**********
esttab model_1 model_3  model_5  model_10 model_11  model_13  model_15  model_20 using average.rtf, compress nonotes r2 label eqlabels(none) title (Estimation of the effect of natural disasters on decentralization) mtitles("1"  "3"  "5" "10" "1"  "3"  "5" "10")  nogaps b(a2) se(a2) star(* 0.10 ** 0.05 *** 0.01) sfmt(%9.0fc) addnotes("Robust standard errors, clustered at the country level, are in parentheses. * p < 0.10, ** p < 0.05, *** p < 0.01.; country and year fixed effects are included; country-specific time trend is included") keep (L.z_rowmax_5dis_maj  L.z_rowmax_5dis_maj3 L.z_rowmax_5dis_maj5 L.z_rowmax_5dis_maj10 L.lnpop  L.lngdppc L.polity2) replace

******Appendix 4-2: Results of models with a full list of control variables****
esttab modela_1 modela_3  modela_5  modela_10 modela_11  modela_13  modela_15  modela_20  using averagea.rtf, compress nonotes r2 label eqlabels(none) title (Estimation of the effect of natural disasters on decentralization) mtitles("1"  "3"  "5" "10" "1"  "3"  "5" "10")  nogaps b(a2) se(a2) star(* 0.10 ** 0.05 *** 0.01) sfmt(%9.0fc) addnotes("Robust standard errors, clustered at the country level, are in parentheses. * p < 0.10, ** p < 0.05, *** p < 0.01.; country and year fixed effects are included; country-specific time trend is included") keep (L.z_rowmax_5dis_maj  L.z_rowmax_5dis_maj3 L.z_rowmax_5dis_maj5 L.z_rowmax_5dis_maj10 L.lnpop  L.lngdppc L.polity2 L.input_gini_disp L.input_EFindex L.pop65  L.lninput_resource) replace



**************Estimating interactive effect between disaster frequency and distance***************************



local i = 21
foreach depvar in fiscal administrative {
foreach indepvar in "c.l.z_rowmax_5dis_maj##c.l.c_zmean_distance_5dis_cap"      "c.l.z_rowmax_5dis_maj2##c.l.c_zmean_distance_5dis_cap2"     "c.l.z_rowmax_5dis_maj3##c.l.c_zmean_distance_5dis_cap3"                        "c.l.z_rowmax_5dis_maj4##c.l.c_zmean_distance_5dis_cap4"  "c.l.z_rowmax_5dis_maj5##c.l.c_zmean_distance_5dis_cap5"                        "c.l.z_rowmax_5dis_maj6##c.l.c_zmean_distance_5dis_cap6" "c.l.z_rowmax_5dis_maj7##c.l.c_zmean_distance_5dis_cap7"                        "c.l.z_rowmax_5dis_maj8##c.l.c_zmean_distance_5dis_cap8" "c.l.z_rowmax_5dis_maj9##c.l.c_zmean_distance_5dis_cap9"                        "c.l.z_rowmax_5dis_maj10##c.l.c_zmean_distance_5dis_cap10"{



reghdfe  `depvar' `indepvar' ${controls} i.scode##c.year , absorb(scode year) cluster(scode)
est store model_`i'
est save model_`i', replace

local i = `i'+1

}
}



***Figure 6. Moderating effect of disaster distance******

coefplot model_21 model_22 model_23 model_24 model_25 model_26 model_27 model_28 model_29 model_30  , keep ("cL.z_rowmax_5dis_maj#cL.c_zmean_distance_5dis_cap"  "cL.z_rowmax_5dis_maj2#cL.c_zmean_distance_5dis_cap2"     "cL.z_rowmax_5dis_maj3#cL.c_zmean_distance_5dis_cap3"                        "cL.z_rowmax_5dis_maj4#cL.c_zmean_distance_5dis_cap4"     "cL.z_rowmax_5dis_maj5#cL.c_zmean_distance_5dis_cap5"                        "cL.z_rowmax_5dis_maj6#cL.c_zmean_distance_5dis_cap6"     "cL.z_rowmax_5dis_maj7#cL.c_zmean_distance_5dis_cap7"                        "cL.z_rowmax_5dis_maj8#cL.c_zmean_distance_5dis_cap8"     "cL.z_rowmax_5dis_maj9#cL.c_zmean_distance_5dis_cap9"                        "cL.z_rowmax_5dis_maj10#cL.c_zmean_distance_5dis_cap10") yline(0)  scheme(s2mono) graphregion(fcolor(white) ilcolor(white) lcolor(white)) vertical  title(fiscal) nokey  level(95) coeflabels (cL.z_rowmax_5dis_maj#cL.c_zmean_distance_5dis_cap="1" cL.z_rowmax_5dis_maj2#cL.c_zmean_distance_5dis_cap2="2" cL.z_rowmax_5dis_maj3#cL.c_zmean_distance_5dis_cap3="3" cL.z_rowmax_5dis_maj4#cL.c_zmean_distance_5dis_cap4="4" cL.z_rowmax_5dis_maj5#cL.c_zmean_distance_5dis_cap5="5"  cL.z_rowmax_5dis_maj6#cL.c_zmean_distance_5dis_cap6="6" cL.z_rowmax_5dis_maj7#cL.c_zmean_distance_5dis_cap7="7" cL.z_rowmax_5dis_maj8#cL.c_zmean_distance_5dis_cap8="8" cL.z_rowmax_5dis_maj9#cL.c_zmean_distance_5dis_cap9="9" cL.z_rowmax_5dis_maj10#cL.c_zmean_distance_5dis_cap10="10")

graph save fiscalcountdistanceappend, replace

coefplot model_31 model_32 model_33 model_34 model_35 model_36 model_37 model_38 model_39 model_40  ,keep ("cL.z_rowmax_5dis_maj#cL.c_zmean_distance_5dis_cap"  "cL.z_rowmax_5dis_maj2#cL.c_zmean_distance_5dis_cap2"     "cL.z_rowmax_5dis_maj3#cL.c_zmean_distance_5dis_cap3"                        "cL.z_rowmax_5dis_maj4#cL.c_zmean_distance_5dis_cap4"     "cL.z_rowmax_5dis_maj5#cL.c_zmean_distance_5dis_cap5"                        "cL.z_rowmax_5dis_maj6#cL.c_zmean_distance_5dis_cap6"     "cL.z_rowmax_5dis_maj7#cL.c_zmean_distance_5dis_cap7"                        "cL.z_rowmax_5dis_maj8#cL.c_zmean_distance_5dis_cap8"     "cL.z_rowmax_5dis_maj9#cL.c_zmean_distance_5dis_cap9"                        "cL.z_rowmax_5dis_maj10#cL.c_zmean_distance_5dis_cap10") yline(0)  scheme(s2mono) graphregion(fcolor(white) ilcolor(white) lcolor(white)) vertical  title(administrative) nokey  level(95) coeflabels (cL.z_rowmax_5dis_maj#cL.c_zmean_distance_5dis_cap="1" cL.z_rowmax_5dis_maj2#cL.c_zmean_distance_5dis_cap2="2" cL.z_rowmax_5dis_maj3#cL.c_zmean_distance_5dis_cap3="3" cL.z_rowmax_5dis_maj4#cL.c_zmean_distance_5dis_cap4="4" cL.z_rowmax_5dis_maj5#cL.c_zmean_distance_5dis_cap5="5"  cL.z_rowmax_5dis_maj6#cL.c_zmean_distance_5dis_cap6="6" cL.z_rowmax_5dis_maj7#cL.c_zmean_distance_5dis_cap7="7" cL.z_rowmax_5dis_maj8#cL.c_zmean_distance_5dis_cap8="8" cL.z_rowmax_5dis_maj9#cL.c_zmean_distance_5dis_cap9="9" cL.z_rowmax_5dis_maj10#cL.c_zmean_distance_5dis_cap10="10")

graph save admincountdistanceappend, replace

graph combine fiscalcountdistanceappend.gph admincountdistanceappend.gph , row(1) ycommon scheme(s2mono) graphregion(fcolor(white) ilcolor(white) lcolor(white))


graph save countdistanceeffectappend, replace



********Appendix 5. Estimation of the interactive effect between disaster frequency and distance of a selective set of models*****
esttab model_21 model_23 model_25 model_30 model_31 model_33 model_35 model_40 using interdistance.rtf, compress nonotes r2 label eqlabels(none) title (Estimating the interactive effect between disaster frequency and distance on decentralization) mtitles("yearly" "3 years" "5 years" "10 years" "yearly" "3 years" "5 years" "10 years")  nogaps nogaps b(a2) se(a2) star(* 0.10 ** 0.05 *** 0.01) sfmt(%9.0fc) addnotes("Robust standard errors, clustered at the country level, are in parentheses. * p < 0.10, ** p < 0.05, *** p < 0.01.; country and year fixed effects are included; country-specific time trend is included") keep ("cL.z_rowmax_5dis_maj#cL.c_zmean_distance_5dis_cap" "cL.z_rowmax_5dis_maj3#cL.c_zmean_distance_5dis_cap3"  "cL.z_rowmax_5dis_maj5#cL.c_zmean_distance_5dis_cap5" "cL.z_rowmax_5dis_maj10#cL.c_zmean_distance_5dis_cap10" L.z_rowmax_5dis_maj  L.z_rowmax_5dis_maj3 L.z_rowmax_5dis_maj5 L.z_rowmax_5dis_maj10 L.c_zmean_distance_5dis_cap L.c_zmean_distance_5dis_cap3 L.c_zmean_distance_5dis_cap5 L.c_zmean_distance_5dis_cap10 L.lnpop  L.lngdppc L.polity2 ) replace



***********Figure 7. Marginal effect of disaster frequency on decentralization at all values of disaster distance***************************

gen dis10distance10=l.c.z_rowmax_5dis_maj10*l.c.c_zmean_distance_5dis_cap10

***frequency*distance***
reghdfe  fiscal z_rowmax_5dis_maj10 c_zmean_distance_5dis_cap10 dis10distance10 ${controls1} i.scode##c.year , absorb(scode year) cluster (scode)
do "fiscalcount_distance.do"
graph save marginalfiscalcountdistance, replace


reghdfe  administrative z_rowmax_5dis_maj10 c_zmean_distance_5dis_cap10 dis10distance10 ${controls1} i.scode##c.year , absorb(scode year) cluster (scode)

do "admincount_distance.do"
graph save marginaladmincountdistance, replace

graph combine "marginalfiscalcountdistance.gph" "marginaladmincountdistance.gph" , row(1) scheme(s2mono) 

 


****************Estimating interactive effect between disaster frequency and dispersion********


local i = 41
foreach depvar in fiscal administrative  {
foreach indepvar in "c.l.z_rowmax_5dis_maj##c.l.c_z_dispersion_5dis"       "c.l.z_rowmax_5dis_maj2##c.l.c_z_dispersion_5dis2" "c.l.z_rowmax_5dis_maj3##c.l.c_z_dispersion_5dis3"                           "c.l.z_rowmax_5dis_maj4##c.l.c_z_dispersion_5dis4"  "c.l.z_rowmax_5dis_maj5##c.l.c_z_dispersion_5dis5"                           "c.l.z_rowmax_5dis_maj6##c.l.c_z_dispersion_5dis6" "c.l.z_rowmax_5dis_maj7##c.l.c_z_dispersion_5dis7"                           "c.l.z_rowmax_5dis_maj8##c.l.c_z_dispersion_5dis8"  "c.l.z_rowmax_5dis_maj9##c.l.c_z_dispersion_5dis9"                           "c.l.z_rowmax_5dis_maj10##c.l.c_z_dispersion_5dis10" {



reghdfe  `depvar' `indepvar' ${controls} i.scode##c.year , absorb(scode year) cluster (scode)
est store model_`i'
est save model_`i', replace

local i = `i'+1

}
}


***Figure 8. Moderating effect of disaster dispersion
coefplot model_41 model_42 model_43 model_44 model_45 model_46 model_47 model_48 model_49 model_50 , keep ("cL.z_rowmax_5dis_maj#cL.c_z_dispersion_5dis" "cL.z_rowmax_5dis_maj2#cL.c_z_dispersion_5dis2" "cL.z_rowmax_5dis_maj3#cL.c_z_dispersion_5dis3" "cL.z_rowmax_5dis_maj4#cL.c_z_dispersion_5dis4"  "cL.z_rowmax_5dis_maj5#cL.c_z_dispersion_5dis5" "cL.z_rowmax_5dis_maj6#cL.c_z_dispersion_5dis6" "cL.z_rowmax_5dis_maj7#cL.c_z_dispersion_5dis7"  "cL.z_rowmax_5dis_maj8#cL.c_z_dispersion_5dis8"  "cL.z_rowmax_5dis_maj9#cL.c_z_dispersion_5dis9" "cL.z_rowmax_5dis_maj10#cL.c_z_dispersion_5dis10") yline(0)  scheme(s2mono) graphregion(fcolor(white) ilcolor(white) lcolor(white)) vertical  title(fiscal) nokey coeflabels (cL.z_rowmax_5dis_maj#cL.c_z_dispersion_5dis="1" cL.z_rowmax_5dis_maj2#cL.c_z_dispersion_5dis2="2" cL.z_rowmax_5dis_maj3#cL.c_z_dispersion_5dis3="3" cL.z_rowmax_5dis_maj4#cL.c_z_dispersion_5dis4="4" cL.z_rowmax_5dis_maj5#cL.c_z_dispersion_5dis5="5"  cL.z_rowmax_5dis_maj6#cL.c_z_dispersion_5dis6="6" cL.z_rowmax_5dis_maj7#cL.c_z_dispersion_5dis7="7" cL.z_rowmax_5dis_maj8#cL.c_z_dispersion_5dis8="8" cL.z_rowmax_5dis_maj9#cL.c_z_dispersion_5dis9="9" cL.z_rowmax_5dis_maj10#cL.c_z_dispersion_5dis10="10") level(95)
graph save fiscalcountdispersionappend, replace

coefplot model_51 model_52 model_53 model_54 model_55 model_56 model_57 model_58 model_59 model_60 , keep ("cL.z_rowmax_5dis_maj#cL.c_z_dispersion_5dis" "cL.z_rowmax_5dis_maj2#cL.c_z_dispersion_5dis2" "cL.z_rowmax_5dis_maj3#cL.c_z_dispersion_5dis3" "cL.z_rowmax_5dis_maj4#cL.c_z_dispersion_5dis4"  "cL.z_rowmax_5dis_maj5#cL.c_z_dispersion_5dis5" "cL.z_rowmax_5dis_maj6#cL.c_z_dispersion_5dis6" "cL.z_rowmax_5dis_maj7#cL.c_z_dispersion_5dis7"  "cL.z_rowmax_5dis_maj8#cL.c_z_dispersion_5dis8"  "cL.z_rowmax_5dis_maj9#cL.c_z_dispersion_5dis9" "cL.z_rowmax_5dis_maj10#cL.c_z_dispersion_5dis10") yline(0)  scheme(s2mono) graphregion(fcolor(white) ilcolor(white) lcolor(white)) vertical  title(administrative) nokey coeflabels (cL.z_rowmax_5dis_maj#cL.c_z_dispersion_5dis="1" cL.z_rowmax_5dis_maj2#cL.c_z_dispersion_5dis2="2" cL.z_rowmax_5dis_maj3#cL.c_z_dispersion_5dis3="3" cL.z_rowmax_5dis_maj4#cL.c_z_dispersion_5dis4="4" cL.z_rowmax_5dis_maj5#cL.c_z_dispersion_5dis5="5"  cL.z_rowmax_5dis_maj6#cL.c_z_dispersion_5dis6="6" cL.z_rowmax_5dis_maj7#cL.c_z_dispersion_5dis7="7" cL.z_rowmax_5dis_maj8#cL.c_z_dispersion_5dis8="8" cL.z_rowmax_5dis_maj9#cL.c_z_dispersion_5dis9="9" cL.z_rowmax_5dis_maj10#cL.c_z_dispersion_5dis10="10") level(95)
graph save admincountdispersionappend, replace


graph combine fiscalcountdispersionappend.gph admincountdispersionappend.gph , row(1) ycommon scheme(s2mono) graphregion(fcolor(white) ilcolor(white) lcolor(white))


graph save countdispersioneffectappend, replace


****Appendix 6. Estimation of the interactive effect between disaster frequency and dispersion of a selective set of models******

esttab model_41 model_43 model_45 model_50 model_51 model_53 model_55 model_60 using interdispersion.rtf, compress nonotes r2 label eqlabels(none) title (Estimating the interactive effect between disaster frequency and dispersion on decentralization) mtitles("yearly" "3 years" "5 years" "10 years" "yearly" "3 years" "5 years" "10 years")  nogaps b(a2) se(a2) star(* 0.10 ** 0.05 *** 0.01) sfmt(%9.0fc) addnotes("Robust standard errors, clustered at the country level, are in parentheses. * p < 0.10, ** p < 0.05, *** p < 0.01.; country and year fixed effects are included; country-specific time trend is included") keep ( cL.z_rowmax_5dis_maj#cL.c_z_dispersion_5dis  cL.z_rowmax_5dis_maj3#cL.c_z_dispersion_5dis3  cL.z_rowmax_5dis_maj5#cL.c_z_dispersion_5dis5 cL.z_rowmax_5dis_maj10#cL.c_z_dispersion_5dis10 L.z_rowmax_5dis_maj  L.z_rowmax_5dis_maj3 L.z_rowmax_5dis_maj5 L.z_rowmax_5dis_maj10 L.c_z_dispersion_5dis L.c_z_dispersion_5dis3 L.c_z_dispersion_5dis5 L.c_z_dispersion_5dis10  L.lnpop  L.lngdppc L.polity2) replace
 


***** Figure 9. Marginal effect of disaster frequency on decentralization at all values of disaster dispersion****

gen dis10dispersion10=l.c.z_rowmax_5dis_maj10*l.c.c_z_dispersion_5dis10
reghdfe  fiscal z_rowmax_5dis_maj10 c_z_dispersion_5dis10 dis10dispersion10${controls1} i.scode##c.year , absorb(scode year) cluster (scode)
do "fiscalcount_dispersion.do"
graph save marginalfiscalcountdispersion, replace

reghdfe  administrative z_rowmax_5dis_maj10 c_z_dispersion_5dis10 dis10dispersion10${controls1} i.scode##c.year , absorb(scode year) cluster (scode)


do "admincount_dispersion.do"
graph save marginaladmincountdispersion, replace



graph combine "marginalfiscalcountdispersion.gph" "marginaladmincountdispersion.gph" , row(1) scheme(s2mono) 

graphregion(fcolor(white) ilcolor(white) lcolor(white)) 









***** The effect of disasters on political centralization******

replace auto=. if auto==-999
replace muni=. if muni==-999
replace state=. if state==-999
replace author =. if author ==-999
replace stconst =. if stconst ==-999


egen float federalism = rowtotal(auton muni state stconst)


**average***
local i = 1
foreach depvar in political federalism {
foreach indepvar in l.z_rowmax_5dis_maj l.z_rowmax_5dis_maj2 l.z_rowmax_5dis_maj3 l.z_rowmax_5dis_maj4 l.z_rowmax_5dis_maj5 l.z_rowmax_5dis_maj6 l.z_rowmax_5dis_maj7 l.z_rowmax_5dis_maj8 l.z_rowmax_5dis_maj9 l.z_rowmax_5dis_maj10{



reghdfe  `depvar' `indepvar' ${controls} i.scode##c.year , absorb(scode year) cluster (scode)
est store cmodel_`i'
est save cmodel_`i', replace
local i = `i'+1

}
}


**inter with distance****

local i = 21
foreach depvar in political federalism {
foreach indepvar in "c.l.z_rowmax_5dis_maj##c.l.c_zmean_distance_5dis_cap"  "c.l.z_rowmax_5dis_maj2##c.l.c_zmean_distance_5dis_cap2"     "c.l.z_rowmax_5dis_maj3##c.l.c_zmean_distance_5dis_cap3"                        "c.l.z_rowmax_5dis_maj4##c.l.c_zmean_distance_5dis_cap4"  "c.l.z_rowmax_5dis_maj5##c.l.c_zmean_distance_5dis_cap5"                        "c.l.z_rowmax_5dis_maj6##c.l.c_zmean_distance_5dis_cap6" "c.l.z_rowmax_5dis_maj7##c.l.c_zmean_distance_5dis_cap7"                        "c.l.z_rowmax_5dis_maj8##c.l.c_zmean_distance_5dis_cap8" "c.l.z_rowmax_5dis_maj9##c.l.c_zmean_distance_5dis_cap9"                        "c.l.z_rowmax_5dis_maj10##c.l.c_zmean_distance_5dis_cap10"{



reghdfe  `depvar' `indepvar' ${controls} i.scode##c.year , absorb(scode year) cluster(scode)
est store cmodel_`i'
est save cmodel_`i', replace

local i = `i'+1

}
}


****inter with dispersion*****

local i = 41
foreach depvar in political federalism {
foreach indepvar in "c.l.z_rowmax_5dis_maj##c.l.c_z_dispersion_5dis"    "c.l.z_rowmax_5dis_maj2##c.l.c_z_dispersion_5dis2" "c.l.z_rowmax_5dis_maj3##c.l.c_z_dispersion_5dis3"                          "c.l.z_rowmax_5dis_maj4##c.l.c_z_dispersion_5dis4"  "c.l.z_rowmax_5dis_maj5##c.l.c_z_dispersion_5dis5"                          "c.l.z_rowmax_5dis_maj6##c.l.c_z_dispersion_5dis6" "c.l.z_rowmax_5dis_maj7##c.l.c_z_dispersion_5dis7"                          "c.l.z_rowmax_5dis_maj8##c.l.c_z_dispersion_5dis8"  "c.l.z_rowmax_5dis_maj9##c.l.c_z_dispersion_5dis9"                          "c.l.z_rowmax_5dis_maj10##c.l.c_z_dispersion_5dis10" {



reghdfe  `depvar' `indepvar' ${controls} i.scode##c.year , absorb(scode year) cluster (scode)
est store cmodel_`i'
est save cmodel_`i', replace

local i = `i'+1

}
}



*****Figure 10. Coefficient plots of the effect of disasters (average effect and interactive effect with distance and dispersion) on political decentralization (RAI)*****
coefplot cmodel_1 cmodel_2 cmodel_3 cmodel_4 cmodel_5 cmodel_6 cmodel_7 cmodel_8 cmodel_9 cmodel_10 , keep (L.z_rowmax_5dis_maj L.z_rowmax_5dis_maj2 L.z_rowmax_5dis_maj3 L.z_rowmax_5dis_maj4 L.z_rowmax_5dis_maj5 L.z_rowmax_5dis_maj6 L.z_rowmax_5dis_maj7 L.z_rowmax_5dis_maj8 L.z_rowmax_5dis_maj9 L.z_rowmax_5dis_maj10) yline(0)  scheme(s2mono) graphregion(fcolor(white) ilcolor(white) lcolor(white)) vertical coeflabels (L.z_rowmax_5dis_maj="1" L.z_rowmax_5dis_maj2="2"  L.z_rowmax_5dis_maj3="3" L.z_rowmax_5dis_maj4="4" L.z_rowmax_5dis_maj5="5" L.z_rowmax_5dis_maj6="6" L.z_rowmax_5dis_maj7="7" L.z_rowmax_5dis_maj8="8" L.z_rowmax_5dis_maj9="9"  L.z_rowmax_5dis_maj10="10" ) title(average effect) nokey
graph save politicalcountappend, replace



coefplot cmodel_21 cmodel_22 cmodel_23 cmodel_24 cmodel_25 cmodel_26 cmodel_27 cmodel_28 cmodel_29 cmodel_30  , keep ("cL.z_rowmax_5dis_maj#cL.c_zmean_distance_5dis_cap"  "cL.z_rowmax_5dis_maj2#cL.c_zmean_distance_5dis_cap2"     "cL.z_rowmax_5dis_maj3#cL.c_zmean_distance_5dis_cap3"                        "cL.z_rowmax_5dis_maj4#cL.c_zmean_distance_5dis_cap4"  "cL.z_rowmax_5dis_maj5#cL.c_zmean_distance_5dis_cap5"                        "cL.z_rowmax_5dis_maj6#cL.c_zmean_distance_5dis_cap6" "cL.z_rowmax_5dis_maj7#cL.c_zmean_distance_5dis_cap7"                        "cL.z_rowmax_5dis_maj8#cL.c_zmean_distance_5dis_cap8" "cL.z_rowmax_5dis_maj9#cL.c_zmean_distance_5dis_cap9"                        "cL.z_rowmax_5dis_maj10#cL.c_zmean_distance_5dis_cap10") yline(0)  scheme(s2mono) graphregion(fcolor(white) ilcolor(white) lcolor(white)) vertical   nokey  level(95) coeflabels (cL.z_rowmax_5dis_maj#cL.c_zmean_distance_5dis_cap="1" cL.z_rowmax_5dis_maj2#cL.c_zmean_distance_5dis_cap2="2" cL.z_rowmax_5dis_maj3#cL.c_zmean_distance_5dis_cap3="3" cL.z_rowmax_5dis_maj4#cL.c_zmean_distance_5dis_cap4="4" cL.z_rowmax_5dis_maj5#cL.c_zmean_distance_5dis_cap5="5"  cL.z_rowmax_5dis_maj6#cL.c_zmean_distance_5dis_cap6="6" cL.z_rowmax_5dis_maj7#cL.c_zmean_distance_5dis_cap7="7" cL.z_rowmax_5dis_maj8#cL.c_zmean_distance_5dis_cap8="8" cL.z_rowmax_5dis_maj9#cL.c_zmean_distance_5dis_cap9="9" cL.z_rowmax_5dis_maj10#cL.c_zmean_distance_5dis_cap10="10") title(frequency*distance)

graph save politicalcountdistanceappend, replace

coefplot cmodel_41 cmodel_42 cmodel_43 cmodel_44 cmodel_45 cmodel_46 cmodel_47 cmodel_48 cmodel_49 cmodel_50  , keep ("cL.z_rowmax_5dis_maj#cL.c_z_dispersion_5dis" "cL.z_rowmax_5dis_maj2#cL.c_z_dispersion_5dis2" "cL.z_rowmax_5dis_maj3#cL.c_z_dispersion_5dis3" "cL.z_rowmax_5dis_maj4#cL.c_z_dispersion_5dis4"  "cL.z_rowmax_5dis_maj5#cL.c_z_dispersion_5dis5" "cL.z_rowmax_5dis_maj6#cL.c_z_dispersion_5dis6" "cL.z_rowmax_5dis_maj7#cL.c_z_dispersion_5dis7"  "cL.z_rowmax_5dis_maj8#cL.c_z_dispersion_5dis8"  "cL.z_rowmax_5dis_maj9#cL.c_z_dispersion_5dis9" "cL.z_rowmax_5dis_maj10#cL.c_z_dispersion_5dis10") yline(0)  scheme(s2mono) graphregion(fcolor(white) ilcolor(white) lcolor(white)) vertical   nokey coeflabels (cL.z_rowmax_5dis_maj#cL.c_z_dispersion_5dis="1" cL.z_rowmax_5dis_maj2#cL.c_z_dispersion_5dis2="2" cL.z_rowmax_5dis_maj3#cL.c_z_dispersion_5dis3="3" cL.z_rowmax_5dis_maj4#cL.c_z_dispersion_5dis4="4" cL.z_rowmax_5dis_maj5#cL.c_z_dispersion_5dis5="5"  cL.z_rowmax_5dis_maj6#cL.c_z_dispersion_5dis6="6" cL.z_rowmax_5dis_maj7#cL.c_z_dispersion_5dis7="7" cL.z_rowmax_5dis_maj8#cL.c_z_dispersion_5dis8="8" cL.z_rowmax_5dis_maj9#cL.c_z_dispersion_5dis9="9" cL.z_rowmax_5dis_maj10#cL.c_z_dispersion_5dis10="10") level(95) title(frequency* dispersion)
graph save politicalcountdispersionappend, replace



graph combine politicalcountappend.gph politicalcountdistanceappend.gph politicalcountdispersionappend.gph  , row(1) ycommon scheme(s2mono) graphregion(fcolor(white) ilcolor(white) lcolor(white)) title (political decentralization_RAI)

graph save politicalappend, replace 


*****Appendix 7****Coefficient plots of the effect of disasters on political centralization as measured in DPI******
coefplot cmodel_11 cmodel_12 cmodel_13 cmodel_14 cmodel_15 cmodel_16 cmodel_17 cmodel_18 cmodel_19 cmodel_20 ,  keep (L.z_rowmax_5dis_maj L.z_rowmax_5dis_maj2 L.z_rowmax_5dis_maj3 L.z_rowmax_5dis_maj4 L.z_rowmax_5dis_maj5 L.z_rowmax_5dis_maj6 L.z_rowmax_5dis_maj7 L.z_rowmax_5dis_maj8 L.z_rowmax_5dis_maj9 L.z_rowmax_5dis_maj10) yline(0)  scheme(s2mono) graphregion(fcolor(white) ilcolor(white) lcolor(white)) vertical coeflabels (L.z_rowmax_5dis_maj="1" L.z_rowmax_5dis_maj2="2"  L.z_rowmax_5dis_maj3="3" L.z_rowmax_5dis_maj4="4" L.z_rowmax_5dis_maj5="5" L.z_rowmax_5dis_maj6="6" L.z_rowmax_5dis_maj7="7" L.z_rowmax_5dis_maj8="8" L.z_rowmax_5dis_maj9="9"  L.z_rowmax_5dis_maj10="10" ) title(average effect) nokey
graph save politicalcountappendf, replace




coefplot cmodel_31 cmodel_32 cmodel_33 cmodel_34 cmodel_35 cmodel_36 cmodel_37 cmodel_38 cmodel_39 cmodel_40  , keep ("cL.z_rowmax_5dis_maj#cL.c_zmean_distance_5dis_cap"  "cL.z_rowmax_5dis_maj2#cL.c_zmean_distance_5dis_cap2"     "cL.z_rowmax_5dis_maj3#cL.c_zmean_distance_5dis_cap3"                        "cL.z_rowmax_5dis_maj4#cL.c_zmean_distance_5dis_cap4"  "cL.z_rowmax_5dis_maj5#cL.c_zmean_distance_5dis_cap5"                        "cL.z_rowmax_5dis_maj6#cL.c_zmean_distance_5dis_cap6" "cL.z_rowmax_5dis_maj7#cL.c_zmean_distance_5dis_cap7"                        "cL.z_rowmax_5dis_maj8#cL.c_zmean_distance_5dis_cap8" "cL.z_rowmax_5dis_maj9#cL.c_zmean_distance_5dis_cap9"                        "cL.z_rowmax_5dis_maj10#cL.c_zmean_distance_5dis_cap10") yline(0)  scheme(s2mono) graphregion(fcolor(white) ilcolor(white) lcolor(white)) vertical   nokey  level(95) coeflabels (cL.z_rowmax_5dis_maj#cL.c_zmean_distance_5dis_cap="1" cL.z_rowmax_5dis_maj2#cL.c_zmean_distance_5dis_cap2="2" cL.z_rowmax_5dis_maj3#cL.c_zmean_distance_5dis_cap3="3" cL.z_rowmax_5dis_maj4#cL.c_zmean_distance_5dis_cap4="4" cL.z_rowmax_5dis_maj5#cL.c_zmean_distance_5dis_cap5="5"  cL.z_rowmax_5dis_maj6#cL.c_zmean_distance_5dis_cap6="6" cL.z_rowmax_5dis_maj7#cL.c_zmean_distance_5dis_cap7="7" cL.z_rowmax_5dis_maj8#cL.c_zmean_distance_5dis_cap8="8" cL.z_rowmax_5dis_maj9#cL.c_zmean_distance_5dis_cap9="9" cL.z_rowmax_5dis_maj10#cL.c_zmean_distance_5dis_cap10="10") title(frequency*distance)

graph save politicalcountdistanceappendf, replace



coefplot cmodel_51 cmodel_52 cmodel_53 cmodel_54 cmodel_55 cmodel_56 cmodel_57 cmodel_58 cmodel_59 cmodel_60  , keep ("cL.z_rowmax_5dis_maj#cL.c_z_dispersion_5dis" "cL.z_rowmax_5dis_maj2#cL.c_z_dispersion_5dis2" "cL.z_rowmax_5dis_maj3#cL.c_z_dispersion_5dis3" "cL.z_rowmax_5dis_maj4#cL.c_z_dispersion_5dis4"  "cL.z_rowmax_5dis_maj5#cL.c_z_dispersion_5dis5" "cL.z_rowmax_5dis_maj6#cL.c_z_dispersion_5dis6" "cL.z_rowmax_5dis_maj7#cL.c_z_dispersion_5dis7"  "cL.z_rowmax_5dis_maj8#cL.c_z_dispersion_5dis8"  "cL.z_rowmax_5dis_maj9#cL.c_z_dispersion_5dis9" "cL.z_rowmax_5dis_maj10#cL.c_z_dispersion_5dis10") yline(0)  scheme(s2mono) graphregion(fcolor(white) ilcolor(white) lcolor(white)) vertical   nokey coeflabels (cL.z_rowmax_5dis_maj#cL.c_z_dispersion_5dis="1" cL.z_rowmax_5dis_maj2#cL.c_z_dispersion_5dis2="2" cL.z_rowmax_5dis_maj3#cL.c_z_dispersion_5dis3="3" cL.z_rowmax_5dis_maj4#cL.c_z_dispersion_5dis4="4" cL.z_rowmax_5dis_maj5#cL.c_z_dispersion_5dis5="5"  cL.z_rowmax_5dis_maj6#cL.c_z_dispersion_5dis6="6" cL.z_rowmax_5dis_maj7#cL.c_z_dispersion_5dis7="7" cL.z_rowmax_5dis_maj8#cL.c_z_dispersion_5dis8="8" cL.z_rowmax_5dis_maj9#cL.c_z_dispersion_5dis9="9" cL.z_rowmax_5dis_maj10#cL.c_z_dispersion_5dis10="10") level(95) title(frequency* dispersion)
graph save politicalcountdispersionappendf, replace



graph combine politicalcountappendf.gph politicalcountdistanceappendf.gph politicalcountdispersionappendf.gph  , row(1) ycommon scheme(s2mono) graphregion(fcolor(white) ilcolor(white) lcolor(white)) title (political decentralization_DPI)

graph save politicalappendf, replace  

******7-2 RAI regression coefficients

esttab cmodel_1  cmodel_3 cmodel_5  cmodel_10 cmodel_21  cmodel_23 cmodel_25 cmodel_30 cmodel_41 cmodel_43  cmodel_45 cmodel_50  using politicalRAI.rtf, compress nonotes r2 label eqlabels(none) title (Regression coefficients: the effect of disasters on political centralization as measured in RAI) mtitles("yearly" "3 years" "5 years" "10 years" "yearly" "3 years" "5 years" "10 years" "yearly" "3 years" "5 years" "10 years" )  nogaps b(a2) se(a2) star(* 0.10 ** 0.05 *** 0.01) sfmt(%9.0fc) addnotes("Robust standard errors, clustered at the country level, are in parentheses. * p < 0.10, ** p < 0.05, *** p < 0.01.; country and year fixed effects are included; country-specific time trend is included") keep ( L.z_rowmax_5dis_maj  L.z_rowmax_5dis_maj3 L.z_rowmax_5dis_maj5 L.z_rowmax_5dis_maj10 L.c_zmean_distance_5dis_cap L.c_zmean_distance_5dis_cap3 L.c_zmean_distance_5dis_cap5 L.c_zmean_distance_5dis_cap10 cL.z_rowmax_5dis_maj#cL.c_zmean_distance_5dis_cap cL.z_rowmax_5dis_maj3#cL.c_zmean_distance_5dis_cap3  cL.z_rowmax_5dis_maj5#cL.c_zmean_distance_5dis_cap5 cL.z_rowmax_5dis_maj10#cL.c_zmean_distance_5dis_cap10 L.c_z_dispersion_5dis L.c_z_dispersion_5dis3 L.c_z_dispersion_5dis5 L.c_z_dispersion_5dis10  cL.z_rowmax_5dis_maj#cL.c_z_dispersion_5dis  cL.z_rowmax_5dis_maj3#cL.c_z_dispersion_5dis3  cL.z_rowmax_5dis_maj5#cL.c_z_dispersion_5dis5 cL.z_rowmax_5dis_maj10#cL.c_z_dispersion_5dis10  L.lnpop  L.lngdppc L.polity2) replace




******7-3 DPI regression coefficients
esttab cmodel_11  cmodel_13 cmodel_15  cmodel_20 cmodel_31  cmodel_33 cmodel_35 cmodel_40 cmodel_51 cmodel_53  cmodel_55 cmodel_60  using politicalDPI.rtf, compress nonotes r2 label eqlabels(none) title (Regression coefficients: the effect of disasters on political centralization as measured in DPI) mtitles("yearly" "3 years" "5 years" "10 years" "yearly" "3 years" "5 years" "10 years" "yearly" "3 years" "5 years" "10 years" )  nogaps b(a2) se(a2) star(* 0.10 ** 0.05 *** 0.01) sfmt(%9.0fc) addnotes("Robust standard errors, clustered at the country level, are in parentheses. * p < 0.10, ** p < 0.05, *** p < 0.01.; country and year fixed effects are included; country-specific time trend is included") keep ( L.z_rowmax_5dis_maj  L.z_rowmax_5dis_maj3 L.z_rowmax_5dis_maj5 L.z_rowmax_5dis_maj10 L.c_zmean_distance_5dis_cap L.c_zmean_distance_5dis_cap3 L.c_zmean_distance_5dis_cap5 L.c_zmean_distance_5dis_cap10 cL.z_rowmax_5dis_maj#cL.c_zmean_distance_5dis_cap cL.z_rowmax_5dis_maj3#cL.c_zmean_distance_5dis_cap3  cL.z_rowmax_5dis_maj5#cL.c_zmean_distance_5dis_cap5 cL.z_rowmax_5dis_maj10#cL.c_zmean_distance_5dis_cap10 L.c_z_dispersion_5dis L.c_z_dispersion_5dis3 L.c_z_dispersion_5dis5 L.c_z_dispersion_5dis10  cL.z_rowmax_5dis_maj#cL.c_z_dispersion_5dis  cL.z_rowmax_5dis_maj3#cL.c_z_dispersion_5dis3  cL.z_rowmax_5dis_maj5#cL.c_z_dispersion_5dis5 cL.z_rowmax_5dis_maj10#cL.c_z_dispersion_5dis10  L.lnpop  L.lngdppc L.polity2) replace





*******************************************different types of disasters***************************************


gen disastertypeA= row_max_n_earth_mag5+ row_max_n_storm+ row_max_n_vol
gen disastertypeA2=row_max_n_earth_mag52+ row_max_n_storm2+ row_max_n_vol2
gen disastertypeA3= row_max_n_earth_mag53+ row_max_n_storm3+ row_max_n_vol3
gen disastertypeA4= row_max_n_earth_mag54+ row_max_n_storm4+ row_max_n_vol4
gen disastertypeA5= row_max_n_earth_mag55+ row_max_n_storm5+ row_max_n_vol5
gen disastertypeA6= row_max_n_earth_mag56+ row_max_n_storm6+ row_max_n_vol6
gen disastertypeA7= row_max_n_earth_mag57+ row_max_n_storm7+ row_max_n_vol7
gen disastertypeA8= row_max_n_earth_mag58+ row_max_n_storm8+ row_max_n_vol8
gen disastertypeA9= row_max_n_earth_mag59+ row_max_n_storm9+ row_max_n_vol9
gen disastertypeA10= row_max_n_earth_mag510+ row_max_n_storm10+ row_max_n_vol10
gen disastertypeB=row_max_n_spi_dr+row_max_n_spi_fl
gen disastertypeB2=row_max_n_spi_dr2+row_max_n_spi_fl2
gen disastertypeB3=row_max_n_spi_dr3+row_max_n_spi_fl3
gen disastertypeB4=row_max_n_spi_dr4+row_max_n_spi_fl4
gen disastertypeB5=row_max_n_spi_dr5+row_max_n_spi_fl5
gen disastertypeB6=row_max_n_spi_dr6+row_max_n_spi_fl6
gen disastertypeB7=row_max_n_spi_dr7+row_max_n_spi_fl7
gen disastertypeB8=row_max_n_spi_dr8+row_max_n_spi_fl8
gen disastertypeB9=row_max_n_spi_dr9+row_max_n_spi_fl9
gen disastertypeB10=row_max_n_spi_dr10+row_max_n_spi_fl10

**type A***

local i = 1
foreach depvar in  fiscal administrative  {
foreach indepvar in l.disastertypeA l.disastertypeA2 l.disastertypeA3 l.disastertypeA4 l.disastertypeA5 l.disastertypeA6 l.disastertypeA7 l.disastertypeA8 l.disastertypeA9 l.disastertypeA10  {



reghdfe   `depvar' `indepvar'  ${controls} i.scode##c.year , absorb(scode year) cluster (scode)
est store modelA_`i'

local i = `i'+1

}
}



**type B***

local i = 1
foreach depvar in  fiscal administrative  {
foreach indepvar in l.disastertypeB l.disastertypeB2 l.disastertypeB3 l.disastertypeB4 l.disastertypeB5 l.disastertypeB6 l.disastertypeB7 l.disastertypeB8 l.disastertypeB9 l.disastertypeB10   {



reghdfe   `depvar' `indepvar'  ${controls} i.scode##c.year , absorb(scode year) cluster (scode)
est store modelB_`i'

local i = `i'+1

}
}

*****Figure 11. Effect of different types of disasters on institutional centralization***********
coefplot modelA_1 modelA_2 modelA_3 modelA_4 modelA_5 modelA_6 modelA_7 modelA_8 modelA_9 modelA_10, keep (L.disastertypeA L.disastertypeA2 L.disastertypeA3 L.disastertypeA4 L.disastertypeA5 L.disastertypeA6 L.disastertypeA7 L.disastertypeA8 L.disastertypeA9 L.disastertypeA10) yline(0) scheme(s2mono) graphregion(fcolor(white) ilcolor(white) lcolor(white)) vertical coeflabels(L.disastertypeA="1" L.disastertypeA2="2" L.disastertypeA3="3" L.disastertypeA4="4" L.disastertypeA5="5" L.disastertypeA6="6" L.disastertypeA7="7"  L.disastertypeA8="8" L.disastertypeA9="9"  L.disastertypeA10="10" ) title(fiscal)  nokey
graph save fiscaltypeAcount, replace

coefplot modelA_11 modelA_12 modelA_13 modelA_14 modelA_15 modelA_16 modelA_17 modelA_18 modelA_19 modelA_20 , keep (L.disastertypeA L.disastertypeA2 L.disastertypeA3 L.disastertypeA4 L.disastertypeA5 L.disastertypeA6 L.disastertypeA7 L.disastertypeA8 L.disastertypeA9 L.disastertypeA10) yline(0) scheme(s2mono) graphregion(fcolor(white) ilcolor(white) lcolor(white)) vertical coeflabels(L.disastertypeA="1" L.disastertypeA2="2" L.disastertypeA3="3" L.disastertypeA4="4" L.disastertypeA5="5" L.disastertypeA6="6" L.disastertypeA7="7"  L.disastertypeA8="8" L.disastertypeA9="9"  L.disastertypeA10="10" )  title(administrative)  nokey
graph save admintypeAcount, replace

graph combine fiscaltypeAcount.gph admintypeAcount.gph  , row(1) ycommon scheme(s2mono) graphregion(fcolor(white) ilcolor(white) lcolor(white)) title("earthquakes, storms, volcanoes")

graph save typeBcount, replace


graph save typeAcount, replace
coefplot modelB_1 modelB_2 modelB_3 modelB_4  modelB_5 modelB_6 modelB_7 modelB_8  modelB_9 modelB_10, keep (L.disastertypeB L.disastertypeB2 L.disastertypeB3 L.disastertypeB4 L.disastertypeB5 L.disastertypeB6 L.disastertypeB7 L.disastertypeB8 L.disastertypeB9 L.disastertypeB10) yline(0) scheme(s2mono) graphregion(fcolor(white) ilcolor(white) lcolor(white)) vertical coeflabels(L.disastertypeB="1" L.disastertypeB2="2" L.disastertypeB3="3" L.disastertypeB4="4" L.disastertypeB5="5" L.disastertypeB6="6" L.disastertypeB7="7"  L.disastertypeB8="8" L.disastertypeB9="9"  L.disastertypeB10="10" )  title(fiscal)  nokey
graph save fiscaltypeBcount, replace 
coefplot modelB_11 modelB_12 modelB_13 modelB_14  modelB_15 modelB_16 modelB_17 modelB_18  modelB_19 modelB_20, keep (L.disastertypeB L.disastertypeB2 L.disastertypeB3 L.disastertypeB4 L.disastertypeB5 L.disastertypeB6 L.disastertypeB7 L.disastertypeB8 L.disastertypeB9 L.disastertypeB10) yline(0) scheme(s2mono) graphregion(fcolor(white) ilcolor(white) lcolor(white)) vertical coeflabels(L.disastertypeB="1" L.disastertypeB2="2" L.disastertypeB3="3" L.disastertypeB4="4" L.disastertypeB5="5" L.disastertypeB6="6" L.disastertypeB7="7"  L.disastertypeB8="8" L.disastertypeB9="9"  L.disastertypeB10="10" ) title(administrative)  nokey
graph save admintypeBcount, replace 

graph combine fiscaltypeBcount.gph admintypeBcount.gph , row(1) ycommon scheme(s2mono) graphregion(fcolor(white) ilcolor(white) lcolor(white)) title("floods, droughts")


graph save typeBcount, replace


graph combine typeAcount.gph typeBcount.gph  , row(2) ycommon scheme(s2mono) graphregion(fcolor(white) ilcolor(white) lcolor(white))



graph save typescount, replace


*****Appendix 8. Additional analyses: varying the measurement of natural disasters
***8-1. Regression coefficient: the effect of two types of disasters
******

esttab modelA_1  modelA_3  modelA_5 modelA_10 modelA_11  modelA_13  modelA_15  modelA_20 modelB_1  modelB_3  modelB_5 modelB_10 modelB_11  modelB_13  modelB_15  modelB_20 using twoTypes.rtf, compress nonotes r2 label eqlabels(none) title ( Regression coefficient: the effect of two types of disasters) mtitles("yearly" "3 years" "5 years" "10 years" "yearly" "3 years" "5 years" "10 years" "yearly" "3 years" "5 years" "10 years" "yearly" "3 years" "5 years" "10 years" "yearly" "3 years" "5 years" "10 years" "yearly" "3 years" "5 years" "10 years" )  nogaps b(a2) se(a2) star(* 0.10 ** 0.05 *** 0.01) sfmt(%9.0fc) addnotes("Robust standard errors, clustered at the country level, are in parentheses. * p < 0.10, ** p < 0.05, *** p < 0.01.; country and year fixed effects are included; country-specific time trend is included") keep ( L.disastertypeA  L.disastertypeA3 L.disastertypeA5  L.disastertypeA10  L.disastertypeB  L.disastertypeB3 L.disastertypeB5  L.disastertypeB10 L.lnpop  L.lngdppc L.polity2) replace






*****add wildfire**** 

*****Appendix 8-2. Coefficient plots showing the effect of disasters that include wildfires on institutional centralization****

local i = 1
foreach depvar in fiscal administrative  {
foreach indepvar in l.z_rowmax_6dis_maj l.z_rowmax_6dis_maj2 l.z_rowmax_6dis_maj3 l.z_rowmax_6dis_maj4  l.z_rowmax_6dis_maj5 l.z_rowmax_6dis_maj6 l.z_rowmax_6dis_maj7 l.z_rowmax_6dis_maj8 l.z_rowmax_6dis_maj9 l.z_rowmax_6dis_maj10{



reghdfe  `depvar' `indepvar' ${controls} i.scode##c.year , absorb(scode year) cluster (scode)
est store wmodel_`i'
est save wmodel_`i', replace
local i = `i'+1

}
}


coefplot wmodel_1 wmodel_2 wmodel_3 wmodel_4 wmodel_5 wmodel_6 wmodel_7 wmodel_8 wmodel_9 wmodel_10 , keep (L.z_rowmax_6dis_maj L.z_rowmax_6dis_maj2 L.z_rowmax_6dis_maj3 L.z_rowmax_6dis_maj4  L.z_rowmax_6dis_maj5 L.z_rowmax_6dis_maj6 L.z_rowmax_6dis_maj7 L.z_rowmax_6dis_maj8 L.z_rowmax_6dis_maj9 L.z_rowmax_6dis_maj10) yline(0)  scheme(s2mono) graphregion(fcolor(white) ilcolor(white) lcolor(white)) vertical coeflabels (L.z_rowmax_6dis_maj="1" L.z_rowmax_6dis_maj2="2"  L.z_rowmax_6dis_maj3="3" L.z_rowmax_6dis_maj4="4" L.z_rowmax_6dis_maj5="5" L.z_rowmax_6dis_maj6="6" L.z_rowmax_6dis_maj7="7" L.z_rowmax_6dis_maj8="8" L.z_rowmax_6dis_maj9="9"  L.z_rowmax_6dis_maj10="10" ) title(fiscal) nokey
graph save fiscalcountappendw, replace

coefplot wmodel_11 wmodel_12 wmodel_13 wmodel_14 wmodel_15 wmodel_16 wmodel_17 wmodel_18 wmodel_19 wmodel_20,  keep (L.z_rowmax_6dis_maj L.z_rowmax_6dis_maj2 L.z_rowmax_6dis_maj3 L.z_rowmax_6dis_maj4  L.z_rowmax_6dis_maj5 L.z_rowmax_6dis_maj6 L.z_rowmax_6dis_maj7 L.z_rowmax_6dis_maj8 L.z_rowmax_6dis_maj9 L.z_rowmax_6dis_maj10) yline(0)  scheme(s2mono) graphregion(fcolor(white) ilcolor(white) lcolor(white)) vertical coeflabels (L.z_rowmax_6dis_maj="1" L.z_rowmax_6dis_maj2="2"  L.z_rowmax_6dis_maj3="3" L.z_rowmax_6dis_maj4="4" L.z_rowmax_6dis_maj5="5" L.z_rowmax_6dis_maj6="6" L.z_rowmax_6dis_maj7="7" L.z_rowmax_6dis_maj8="8" L.z_rowmax_6dis_maj9="9"  L.z_rowmax_6dis_maj10="10"  ) title(administrative) nokey
graph save wadmincountappend, replace



graph combine fiscalcountappendw.gph wadmincountappend.gph , row(2) ycommon scheme(s2mono) graphregion(fcolor(white) ilcolor(white) lcolor(white))


graph save counteffectappendw, replace

***********8-3. Regression coefficient: the effect of disasters that include wildfires on institutional centralization***********
esttab wmodel_1  wmodel_3 wmodel_5 wmodel_10 wmodel_11  wmodel_13 wmodel_15 wmodel_20 using wildfire.rtf, compress nonotes r2 label eqlabels(none)  mtitles("yearly" "3 years" "5 years" "10 years" "yearly" "3 years" "5 years" "10 years" )  nogaps b(a2) se(a2) star(* 0.10 ** 0.05 *** 0.01) sfmt(%9.0fc) addnotes("Robust standard errors, clustered at the country level, are in parentheses. * p < 0.10, ** p < 0.05, *** p < 0.01.; country and year fixed effects are included; country-specific time trend is included") keep ( L.z_rowmax_6dis_maj  L.z_rowmax_6dis_maj3   L.z_rowmax_6dis_maj5 L.z_rowmax_6dis_maj10  L.lnpop  L.lngdppc L.polity2) replace




******Appendix 9, the effect of distance to most populuous cities *****


local i = 21
foreach depvar in fiscal administrative {
foreach indepvar in "c.l.z_rowmax_5dis_maj##c.l.c_zmean_distance_5dis_big"  "c.l.z_rowmax_5dis_maj2##c.l.c_zmean_distance_5dis_big2"     "c.l.z_rowmax_5dis_maj3##c.l.c_zmean_distance_5dis_big3"                        "c.l.z_rowmax_5dis_maj4##c.l.c_zmean_distance_5dis_big4"  "c.l.z_rowmax_5dis_maj5##c.l.c_zmean_distance_5dis_big5"                        "c.l.z_rowmax_5dis_maj6##c.l.c_zmean_distance_5dis_big6" "c.l.z_rowmax_5dis_maj7##c.l.c_zmean_distance_5dis_big7"                        "c.l.z_rowmax_5dis_maj8##c.l.c_zmean_distance_5dis_big8" "c.l.z_rowmax_5dis_maj9##c.l.c_zmean_distance_5dis_big9"                        "c.l.z_rowmax_5dis_maj10##c.l.c_zmean_distance_5dis_big10"{



reghdfe  `depvar' `indepvar' ${controls} i.scode##c.year , absorb(scode year) cluster(scode)
est store modelpop_`i'
est save modelpop_`i', replace

local i = `i'+1

}
}


****Appendix 9-1. Coefficient plots of the interactive effect between disaster frequency and distance to the most populous city on decentralization***** 
coefplot modelpop_21 modelpop_22 modelpop_23 modelpop_24 modelpop_25 modelpop_26 modelpop_27 modelpop_28 modelpop_29 modelpop_30  , keep ("cL.z_rowmax_5dis_maj#cL.c_zmean_distance_5dis_big"                        "cL.z_rowmax_5dis_maj2#cL.c_zmean_distance_5dis_big2"     "cL.z_rowmax_5dis_maj3#cL.c_zmean_distance_5dis_big3"                      "cL.z_rowmax_5dis_maj4#cL.c_zmean_distance_5dis_big4"  "cL.z_rowmax_5dis_maj5#cL.c_zmean_distance_5dis_big5"                      "cL.z_rowmax_5dis_maj6#cL.c_zmean_distance_5dis_big6" "cL.z_rowmax_5dis_maj7#cL.c_zmean_distance_5dis_big7"                      "cL.z_rowmax_5dis_maj8#cL.c_zmean_distance_5dis_big8" "cL.z_rowmax_5dis_maj9#cL.c_zmean_distance_5dis_big9"                      "cL.z_rowmax_5dis_maj10#cL.c_zmean_distance_5dis_big10") yline(0)  scheme(s2mono) graphregion(fcolor(white) ilcolor(white) lcolor(white)) vertical  nokey  level(95) coeflabels (cL.z_rowmax_5dis_maj#cL.c_zmean_distance_5dis_big="1" cL.z_rowmax_5dis_maj2#cL.c_zmean_distance_5dis_big2="2" cL.z_rowmax_5dis_maj3#cL.c_zmean_distance_5dis_big3="3" cL.z_rowmax_5dis_maj4#cL.c_zmean_distance_5dis_big4="4" cL.z_rowmax_5dis_maj5#cL.c_zmean_distance_5dis_big5="5"  cL.z_rowmax_5dis_maj6#cL.c_zmean_distance_5dis_big6="6" cL.z_rowmax_5dis_maj7#cL.c_zmean_distance_5dis_big7="7" cL.z_rowmax_5dis_maj8#cL.c_zmean_distance_5dis_big8="8" cL.z_rowmax_5dis_maj9#cL.c_zmean_distance_5dis_big9="9" cL.z_rowmax_5dis_maj10#cL.c_zmean_distance_5dis_big10="10") title(fiscal) 

graph save fiscalcountdistanceappendpop, replace



coefplot modelpop_31 modelpop_32 modelpop_33 modelpop_34 modelpop_35 modelpop_36 modelpop_37 modelpop_38 modelpop_39 modelpop_40  , keep ("cL.z_rowmax_5dis_maj#cL.c_zmean_distance_5dis_big"                        "cL.z_rowmax_5dis_maj2#cL.c_zmean_distance_5dis_big2"     "cL.z_rowmax_5dis_maj3#cL.c_zmean_distance_5dis_big3"                      "cL.z_rowmax_5dis_maj4#cL.c_zmean_distance_5dis_big4"  "cL.z_rowmax_5dis_maj5#cL.c_zmean_distance_5dis_big5"                      "cL.z_rowmax_5dis_maj6#cL.c_zmean_distance_5dis_big6" "cL.z_rowmax_5dis_maj7#cL.c_zmean_distance_5dis_big7"                      "cL.z_rowmax_5dis_maj8#cL.c_zmean_distance_5dis_big8" "cL.z_rowmax_5dis_maj9#cL.c_zmean_distance_5dis_big9"                      "cL.z_rowmax_5dis_maj10#cL.c_zmean_distance_5dis_big10") yline(0)  scheme(s2mono) graphregion(fcolor(white) ilcolor(white) lcolor(white)) vertical  nokey  level(95) coeflabels (cL.z_rowmax_5dis_maj#cL.c_zmean_distance_5dis_big="1" cL.z_rowmax_5dis_maj2#cL.c_zmean_distance_5dis_big2="2" cL.z_rowmax_5dis_maj3#cL.c_zmean_distance_5dis_big3="3" cL.z_rowmax_5dis_maj4#cL.c_zmean_distance_5dis_big4="4" cL.z_rowmax_5dis_maj5#cL.c_zmean_distance_5dis_big5="5"  cL.z_rowmax_5dis_maj6#cL.c_zmean_distance_5dis_big6="6" cL.z_rowmax_5dis_maj7#cL.c_zmean_distance_5dis_big7="7" cL.z_rowmax_5dis_maj8#cL.c_zmean_distance_5dis_big8="8" cL.z_rowmax_5dis_maj9#cL.c_zmean_distance_5dis_big9="9" cL.z_rowmax_5dis_maj10#cL.c_zmean_distance_5dis_big10="10") title(administrative) 


graph save admincountdistanceappendpop, replace


graph combine fiscalcountdistanceappendpop.gph admincountdistanceappendpop.gph , row(1) ycommon scheme(s2mono) graphregion(fcolor(white) ilcolor(white) lcolor(white))


graph save countdistanceeffectappendpop, replace

********9-2. Regression coefficients: the interactive effect between disaster frequency and distance to the most populous city on decentralization*****


esttab modelpop_21 modelpop_23  modelpop_25 modelpop_30 modelpop_31 modelpop_33  modelpop_35 modelpop_40 using popcity.rtf, compress nonotes r2 label eqlabels(none)  mtitles("yearly" "3 years" "5 years" "10 years" "yearly" "3 years" "5 years" "10 years" )  nogaps b(a2) se(a2) star(* 0.10 ** 0.05 *** 0.01) sfmt(%9.0fc) addnotes("Robust standard errors, clustered at the country level, are in parentheses. * p < 0.10, ** p < 0.05, *** p < 0.01.; country and year fixed effects are included; country-specific time trend is included") keep ( L.z_rowmax_5dis_maj  L.z_rowmax_5dis_maj3   L.z_rowmax_5dis_maj5 L.z_rowmax_5dis_maj10 L.c_zmean_distance_5dis_big L.c_zmean_distance_5dis_big3 L.c_zmean_distance_5dis_big5 L.c_zmean_distance_5dis_big10 "cL.z_rowmax_5dis_maj#cL.c_zmean_distance_5dis_big"                            "cL.z_rowmax_5dis_maj3#cL.c_zmean_distance_5dis_big3"     "cL.z_rowmax_5dis_maj5#cL.c_zmean_distance_5dis_big5" "cL.z_rowmax_5dis_maj10#cL.c_zmean_distance_5dis_big10"  L.lnpop  L.lngdppc L.polity2) replace





*******control for major cities ****

local i = 41
foreach depvar in fiscal administrative {
foreach indepvar in "c.l.z_rowmax_5dis_maj##c.l.c_zmean_distance_5dis_cap l.zmean_distance_5dis_big"  "c.l.z_rowmax_5dis_maj2##c.l.c_zmean_distance_5dis_cap2 l.zmean_distance_5dis_big2"   "c.l.z_rowmax_5dis_maj3##c.l.c_zmean_distance_5dis_cap3 l.zmean_distance_5dis_big3"                     "c.l.z_rowmax_5dis_maj4##c.l.c_zmean_distance_5dis_cap4 l.zmean_distance_5dis_big4"  "c.l.z_rowmax_5dis_maj5##c.l.c_zmean_distance_5dis_cap5 l.zmean_distance_5dis_big5"                        "c.l.z_rowmax_5dis_maj6##c.l.c_zmean_distance_5dis_cap6 l.zmean_distance_5dis_big6"  "c.l.z_rowmax_5dis_maj7##c.l.c_zmean_distance_5dis_cap7 l.zmean_distance_5dis_big7"                       "c.l.z_rowmax_5dis_maj8##c.l.c_zmean_distance_5dis_cap8 l.zmean_distance_5dis_big8"  "c.l.z_rowmax_5dis_maj9##c.l.c_zmean_distance_5dis_cap9 l.zmean_distance_5dis_big9"                         "c.l.z_rowmax_5dis_maj10##c.l.c_zmean_distance_5dis_cap10 l.zmean_distance_5dis_big10"{



reghdfe  `depvar' `indepvar' ${controls}  i.scode##c.year , absorb(scode year) cluster(scode)
est store modelpop_`i'
est save modelpop_`i', replace

local i = `i'+1

}
}

****Appendix 9-3. Coefficient plots of the interactive effect between disaster frequency and distance to the capital city, controlling for the effect of distance to the most populous city*****

coefplot modelpop_41 modelpop_42 modelpop_43 modelpop_44 modelpop_45 modelpop_46 modelpop_47 modelpop_48 modelpop_49 modelpop_50  , keep ("cL.z_rowmax_5dis_maj#cL.c_zmean_distance_5dis_cap"  "cL.z_rowmax_5dis_maj2#cL.c_zmean_distance_5dis_cap2"     "cL.z_rowmax_5dis_maj3#cL.c_zmean_distance_5dis_cap3"                        "cL.z_rowmax_5dis_maj4#cL.c_zmean_distance_5dis_cap4"     "cL.z_rowmax_5dis_maj5#cL.c_zmean_distance_5dis_cap5"                        "cL.z_rowmax_5dis_maj6#cL.c_zmean_distance_5dis_cap6"     "cL.z_rowmax_5dis_maj7#cL.c_zmean_distance_5dis_cap7"                        "cL.z_rowmax_5dis_maj8#cL.c_zmean_distance_5dis_cap8"     "cL.z_rowmax_5dis_maj9#cL.c_zmean_distance_5dis_cap9"                        "cL.z_rowmax_5dis_maj10#cL.c_zmean_distance_5dis_cap10") yline(0)  scheme(s2mono) graphregion(fcolor(white) ilcolor(white) lcolor(white)) vertical  title(fiscal) nokey  level(95) coeflabels (cL.z_rowmax_5dis_maj#cL.c_zmean_distance_5dis_cap="1" cL.z_rowmax_5dis_maj2#cL.c_zmean_distance_5dis_cap2="2" cL.z_rowmax_5dis_maj3#cL.c_zmean_distance_5dis_cap3="3" cL.z_rowmax_5dis_maj4#cL.c_zmean_distance_5dis_cap4="4" cL.z_rowmax_5dis_maj5#cL.c_zmean_distance_5dis_cap5="5"  cL.z_rowmax_5dis_maj6#cL.c_zmean_distance_5dis_cap6="6" cL.z_rowmax_5dis_maj7#cL.c_zmean_distance_5dis_cap7="7" cL.z_rowmax_5dis_maj8#cL.c_zmean_distance_5dis_cap8="8" cL.z_rowmax_5dis_maj9#cL.c_zmean_distance_5dis_cap9="9" cL.z_rowmax_5dis_maj10#cL.c_zmean_distance_5dis_cap10="10")


graph save fiscalcountdistanceappendpopcontrol, replace



coefplot modelpop_51 modelpop_52 modelpop_53 modelpop_54 modelpop_55 modelpop_56 modelpop_57 modelpop_58 modelpop_59 modelpop_60   ,keep ("cL.z_rowmax_5dis_maj#cL.c_zmean_distance_5dis_cap"  "cL.z_rowmax_5dis_maj2#cL.c_zmean_distance_5dis_cap2"     "cL.z_rowmax_5dis_maj3#cL.c_zmean_distance_5dis_cap3"                        "cL.z_rowmax_5dis_maj4#cL.c_zmean_distance_5dis_cap4"     "cL.z_rowmax_5dis_maj5#cL.c_zmean_distance_5dis_cap5"                        "cL.z_rowmax_5dis_maj6#cL.c_zmean_distance_5dis_cap6"     "cL.z_rowmax_5dis_maj7#cL.c_zmean_distance_5dis_cap7"                        "cL.z_rowmax_5dis_maj8#cL.c_zmean_distance_5dis_cap8"     "cL.z_rowmax_5dis_maj9#cL.c_zmean_distance_5dis_cap9"                        "cL.z_rowmax_5dis_maj10#cL.c_zmean_distance_5dis_cap10") yline(0)  scheme(s2mono) graphregion(fcolor(white) ilcolor(white) lcolor(white)) vertical  title(administrative) nokey  level(95) coeflabels (cL.z_rowmax_5dis_maj#cL.c_zmean_distance_5dis_cap="1" cL.z_rowmax_5dis_maj2#cL.c_zmean_distance_5dis_cap2="2" cL.z_rowmax_5dis_maj3#cL.c_zmean_distance_5dis_cap3="3" cL.z_rowmax_5dis_maj4#cL.c_zmean_distance_5dis_cap4="4" cL.z_rowmax_5dis_maj5#cL.c_zmean_distance_5dis_cap5="5"  cL.z_rowmax_5dis_maj6#cL.c_zmean_distance_5dis_cap6="6" cL.z_rowmax_5dis_maj7#cL.c_zmean_distance_5dis_cap7="7" cL.z_rowmax_5dis_maj8#cL.c_zmean_distance_5dis_cap8="8" cL.z_rowmax_5dis_maj9#cL.c_zmean_distance_5dis_cap9="9" cL.z_rowmax_5dis_maj10#cL.c_zmean_distance_5dis_cap10="10")



graph save admincountdistanceappendpopcontrol, replace


graph combine fiscalcountdistanceappendpopcontrol.gph admincountdistanceappendpopcontrol.gph , row(1) ycommon scheme(s2mono) graphregion(fcolor(white) ilcolor(white) lcolor(white))
graph save controlpop, replace


*****9-4. Regression coefficient: the interactive effect between disaster frequency and distance to the capital city, controlling for the effect of distance to the most populous city*****

esttab modelpop_41 modelpop_43  modelpop_45 modelpop_50 modelpop_51 modelpop_53  modelpop_55 modelpop_60 using popcitycontrol.rtf, compress nonotes r2 label eqlabels(none)  mtitles("yearly" "3 years" "5 years" "10 years" "yearly" "3 years" "5 years" "10 years" )  nogaps b(a2) se(a2) star(* 0.10 ** 0.05 *** 0.01) sfmt(%9.0fc) addnotes("Robust standard errors, clustered at the country level, are in parentheses. * p < 0.10, ** p < 0.05, *** p < 0.01.; country and year fixed effects are included; country-specific time trend is included") keep ( L.z_rowmax_5dis_maj  L.z_rowmax_5dis_maj3   L.z_rowmax_5dis_maj5 L.z_rowmax_5dis_maj10 L.c_zmean_distance_5dis_cap L.c_zmean_distance_5dis_cap3 L.c_zmean_distance_5dis_cap5 L.c_zmean_distance_5dis_cap10 "cL.z_rowmax_5dis_maj#cL.c_zmean_distance_5dis_cap"     "cL.z_rowmax_5dis_maj3#cL.c_zmean_distance_5dis_cap3"    "cL.z_rowmax_5dis_maj5#cL.c_zmean_distance_5dis_cap5"                       "cL.z_rowmax_5dis_maj10#cL.c_zmean_distance_5dis_cap10" L.lnpop  L.lngdppc L.polity2) replace




*********Appendix 10. Heterogeneous moderating effect of disaster distance in countries with different levels of economic development****



********developing*****
local i = 1
foreach depvar in fiscal administrative  {
foreach indepvar in "c.l.z_rowmax_5dis_maj##c.l.c_zmean_distance_5dis_cap"      "c.l.z_rowmax_5dis_maj2##c.l.c_zmean_distance_5dis_cap2"     "c.l.z_rowmax_5dis_maj3##c.l.c_zmean_distance_5dis_cap3"                        "c.l.z_rowmax_5dis_maj4##c.l.c_zmean_distance_5dis_cap4"  "c.l.z_rowmax_5dis_maj5##c.l.c_zmean_distance_5dis_cap5"                        "c.l.z_rowmax_5dis_maj6##c.l.c_zmean_distance_5dis_cap6" "c.l.z_rowmax_5dis_maj7##c.l.c_zmean_distance_5dis_cap7"                        "c.l.z_rowmax_5dis_maj8##c.l.c_zmean_distance_5dis_cap8" "c.l.z_rowmax_5dis_maj9##c.l.c_zmean_distance_5dis_cap9"                        "c.l.z_rowmax_5dis_maj10##c.l.c_zmean_distance_5dis_cap10"{



reghdfe  `depvar' `indepvar' l.lnpop l.polity2 i.scode##c.year if GDPpercapitaconstant2010US<10987, absorb(scode year) cluster (scode)
est store D1model_`i'
est save D1model_`i', replace
local i = `i'+1

}
}


****developed*****
local i = 1
foreach depvar in fiscal administrative  {
foreach indepvar in "c.l.z_rowmax_5dis_maj##c.l.c_zmean_distance_5dis_cap"      "c.l.z_rowmax_5dis_maj2##c.l.c_zmean_distance_5dis_cap2"     "c.l.z_rowmax_5dis_maj3##c.l.c_zmean_distance_5dis_cap3"                        "c.l.z_rowmax_5dis_maj4##c.l.c_zmean_distance_5dis_cap4"  "c.l.z_rowmax_5dis_maj5##c.l.c_zmean_distance_5dis_cap5"                        "c.l.z_rowmax_5dis_maj6##c.l.c_zmean_distance_5dis_cap6" "c.l.z_rowmax_5dis_maj7##c.l.c_zmean_distance_5dis_cap7"                        "c.l.z_rowmax_5dis_maj8##c.l.c_zmean_distance_5dis_cap8" "c.l.z_rowmax_5dis_maj9##c.l.c_zmean_distance_5dis_cap9"                        "c.l.z_rowmax_5dis_maj10##c.l.c_zmean_distance_5dis_cap10"{



reghdfe  `depvar' `indepvar' l.lnpop l.polity2 i.scode##c.year if GDPpercapitaconstant2010US>10987, absorb(scode year) cluster (scode)
est store D2model_`i'
est save D2model_`i', replace
local i = `i'+1

}
}

****Appendix 10-1. Coefficient plots of the interaction term between disaster frequency and distance to the capital 

coefplot D1model_1 D1model_2 D1model_3 D1model_4 D1model_5 D1model_6 D1model_7 D1model_8 D1model_9 D1model_10 ,   keep ("cL.z_rowmax_5dis_maj#cL.c_zmean_distance_5dis_cap"  "cL.z_rowmax_5dis_maj2#cL.c_zmean_distance_5dis_cap2"     "cL.z_rowmax_5dis_maj3#cL.c_zmean_distance_5dis_cap3"                        "cL.z_rowmax_5dis_maj4#cL.c_zmean_distance_5dis_cap4"     "cL.z_rowmax_5dis_maj5#cL.c_zmean_distance_5dis_cap5"                        "cL.z_rowmax_5dis_maj6#cL.c_zmean_distance_5dis_cap6"     "cL.z_rowmax_5dis_maj7#cL.c_zmean_distance_5dis_cap7"                        "cL.z_rowmax_5dis_maj8#cL.c_zmean_distance_5dis_cap8"     "cL.z_rowmax_5dis_maj9#cL.c_zmean_distance_5dis_cap9"                        "cL.z_rowmax_5dis_maj10#cL.c_zmean_distance_5dis_cap10") yline(0)  scheme(s2mono) graphregion(fcolor(white) ilcolor(white) lcolor(white)) vertical level(95) coeflabels (cL.z_rowmax_5dis_maj#cL.c_zmean_distance_5dis_cap="1" cL.z_rowmax_5dis_maj2#cL.c_zmean_distance_5dis_cap2="2" cL.z_rowmax_5dis_maj3#cL.c_zmean_distance_5dis_cap3="3" cL.z_rowmax_5dis_maj4#cL.c_zmean_distance_5dis_cap4="4" cL.z_rowmax_5dis_maj5#cL.c_zmean_distance_5dis_cap5="5"  cL.z_rowmax_5dis_maj6#cL.c_zmean_distance_5dis_cap6="6" cL.z_rowmax_5dis_maj7#cL.c_zmean_distance_5dis_cap7="7" cL.z_rowmax_5dis_maj8#cL.c_zmean_distance_5dis_cap8="8" cL.z_rowmax_5dis_maj9#cL.c_zmean_distance_5dis_cap9="9" cL.z_rowmax_5dis_maj10#cL.c_zmean_distance_5dis_cap10="10") title(fiscal) nokey
graph save D1fiscalcountappenda, replace

coefplot D1model_11 D1model_12 D1model_13 D1model_14 D1model_15 D1model_16 D1model_17 D1model_18 D1model_19 D1model_20 ,  keep ("cL.z_rowmax_5dis_maj#cL.c_zmean_distance_5dis_cap"  "cL.z_rowmax_5dis_maj2#cL.c_zmean_distance_5dis_cap2"     "cL.z_rowmax_5dis_maj3#cL.c_zmean_distance_5dis_cap3"                        "cL.z_rowmax_5dis_maj4#cL.c_zmean_distance_5dis_cap4"     "cL.z_rowmax_5dis_maj5#cL.c_zmean_distance_5dis_cap5"                        "cL.z_rowmax_5dis_maj6#cL.c_zmean_distance_5dis_cap6"     "cL.z_rowmax_5dis_maj7#cL.c_zmean_distance_5dis_cap7"                        "cL.z_rowmax_5dis_maj8#cL.c_zmean_distance_5dis_cap8"     "cL.z_rowmax_5dis_maj9#cL.c_zmean_distance_5dis_cap9"                        "cL.z_rowmax_5dis_maj10#cL.c_zmean_distance_5dis_cap10") yline(0)  scheme(s2mono) graphregion(fcolor(white) ilcolor(white) lcolor(white)) vertical level(95) coeflabels (cL.z_rowmax_5dis_maj#cL.c_zmean_distance_5dis_cap="1" cL.z_rowmax_5dis_maj2#cL.c_zmean_distance_5dis_cap2="2" cL.z_rowmax_5dis_maj3#cL.c_zmean_distance_5dis_cap3="3" cL.z_rowmax_5dis_maj4#cL.c_zmean_distance_5dis_cap4="4" cL.z_rowmax_5dis_maj5#cL.c_zmean_distance_5dis_cap5="5"  cL.z_rowmax_5dis_maj6#cL.c_zmean_distance_5dis_cap6="6" cL.z_rowmax_5dis_maj7#cL.c_zmean_distance_5dis_cap7="7" cL.z_rowmax_5dis_maj8#cL.c_zmean_distance_5dis_cap8="8" cL.z_rowmax_5dis_maj9#cL.c_zmean_distance_5dis_cap9="9" cL.z_rowmax_5dis_maj10#cL.c_zmean_distance_5dis_cap10="10") title(administrative) nokey
graph save D1admincountappenda, replace





coefplot D2model_1 D2model_2 D2model_3 D2model_4 D2model_5 D2model_6 D2model_7 D2model_8 D2model_9 D2model_10 ,   keep ("cL.z_rowmax_5dis_maj#cL.c_zmean_distance_5dis_cap"  "cL.z_rowmax_5dis_maj2#cL.c_zmean_distance_5dis_cap2"     "cL.z_rowmax_5dis_maj3#cL.c_zmean_distance_5dis_cap3"                        "cL.z_rowmax_5dis_maj4#cL.c_zmean_distance_5dis_cap4"     "cL.z_rowmax_5dis_maj5#cL.c_zmean_distance_5dis_cap5"                        "cL.z_rowmax_5dis_maj6#cL.c_zmean_distance_5dis_cap6"     "cL.z_rowmax_5dis_maj7#cL.c_zmean_distance_5dis_cap7"                        "cL.z_rowmax_5dis_maj8#cL.c_zmean_distance_5dis_cap8"     "cL.z_rowmax_5dis_maj9#cL.c_zmean_distance_5dis_cap9"                        "cL.z_rowmax_5dis_maj10#cL.c_zmean_distance_5dis_cap10") yline(0)  scheme(s2mono) graphregion(fcolor(white) ilcolor(white) lcolor(white)) vertical level(95) coeflabels (cL.z_rowmax_5dis_maj#cL.c_zmean_distance_5dis_cap="1" cL.z_rowmax_5dis_maj2#cL.c_zmean_distance_5dis_cap2="2" cL.z_rowmax_5dis_maj3#cL.c_zmean_distance_5dis_cap3="3" cL.z_rowmax_5dis_maj4#cL.c_zmean_distance_5dis_cap4="4" cL.z_rowmax_5dis_maj5#cL.c_zmean_distance_5dis_cap5="5"  cL.z_rowmax_5dis_maj6#cL.c_zmean_distance_5dis_cap6="6" cL.z_rowmax_5dis_maj7#cL.c_zmean_distance_5dis_cap7="7" cL.z_rowmax_5dis_maj8#cL.c_zmean_distance_5dis_cap8="8" cL.z_rowmax_5dis_maj9#cL.c_zmean_distance_5dis_cap9="9" cL.z_rowmax_5dis_maj10#cL.c_zmean_distance_5dis_cap10="10") title(fiscal) nokey
graph save D2fiscalcountappenda, replace

coefplot D2model_11 D2model_12 D2model_13 D2model_14 D2model_15 D2model_16 D2model_17 D2model_18 D2model_19 D2model_20,  keep ("cL.z_rowmax_5dis_maj#cL.c_zmean_distance_5dis_cap"  "cL.z_rowmax_5dis_maj2#cL.c_zmean_distance_5dis_cap2"     "cL.z_rowmax_5dis_maj3#cL.c_zmean_distance_5dis_cap3"                        "cL.z_rowmax_5dis_maj4#cL.c_zmean_distance_5dis_cap4"     "cL.z_rowmax_5dis_maj5#cL.c_zmean_distance_5dis_cap5"                        "cL.z_rowmax_5dis_maj6#cL.c_zmean_distance_5dis_cap6"     "cL.z_rowmax_5dis_maj7#cL.c_zmean_distance_5dis_cap7"                        "cL.z_rowmax_5dis_maj8#cL.c_zmean_distance_5dis_cap8"     "cL.z_rowmax_5dis_maj9#cL.c_zmean_distance_5dis_cap9"                        "cL.z_rowmax_5dis_maj10#cL.c_zmean_distance_5dis_cap10") yline(0)  scheme(s2mono) graphregion(fcolor(white) ilcolor(white) lcolor(white)) vertical level(95) coeflabels (cL.z_rowmax_5dis_maj#cL.c_zmean_distance_5dis_cap="1" cL.z_rowmax_5dis_maj2#cL.c_zmean_distance_5dis_cap2="2" cL.z_rowmax_5dis_maj3#cL.c_zmean_distance_5dis_cap3="3" cL.z_rowmax_5dis_maj4#cL.c_zmean_distance_5dis_cap4="4" cL.z_rowmax_5dis_maj5#cL.c_zmean_distance_5dis_cap5="5"  cL.z_rowmax_5dis_maj6#cL.c_zmean_distance_5dis_cap6="6" cL.z_rowmax_5dis_maj7#cL.c_zmean_distance_5dis_cap7="7" cL.z_rowmax_5dis_maj8#cL.c_zmean_distance_5dis_cap8="8" cL.z_rowmax_5dis_maj9#cL.c_zmean_distance_5dis_cap9="9" cL.z_rowmax_5dis_maj10#cL.c_zmean_distance_5dis_cap10="10") title(administrative) nokey
graph save D2admincountappenda, replace


graph combine D1fiscalcountappenda.gph D1admincountappenda.gph , row(1) ycommon scheme(s2mono) graphregion(fcolor(white) ilcolor(white) lcolor(white)) title (poorer countries)


graph save D1counteffectappenda, replace


graph combine D2fiscalcountappenda.gph D2admincountappenda.gph , row(1) ycommon scheme(s2mono) graphregion(fcolor(white) ilcolor(white) lcolor(white)) title (richer countries)


graph save D2counteffectappenda, replace


graph combine D1counteffectappenda.gph D2counteffectappenda.gph , row(2) ycommon scheme(s2mono) graphregion(fcolor(white) ilcolor(white) lcolor(white))

*******10-2. Regression coefficient: moderating effect of disaster distance in countries with different levels of economic development******

esttab D1model_1  D1model_3  D1model_5  D1model_10  D1model_11 D1model_13 D1model_15 D1model_20 using poor.rtf, compress nonotes r2 label eqlabels(none)  mtitles("yearly" "3 years" "5 years" "10 years" "yearly" "3 years" "5 years" "10 years" )  nogaps b(a2) se(a2) star(* 0.10 ** 0.05 *** 0.01) sfmt(%9.0fc) addnotes("Robust standard errors, clustered at the country level, are in parentheses. * p < 0.10, ** p < 0.05, *** p < 0.01.; country and year fixed effects are included; country-specific time trend is included") keep ( L.z_rowmax_5dis_maj  L.z_rowmax_5dis_maj3   L.z_rowmax_5dis_maj5 L.z_rowmax_5dis_maj10 L.c_zmean_distance_5dis_cap L.c_zmean_distance_5dis_cap3 L.c_zmean_distance_5dis_cap5 L.c_zmean_distance_5dis_cap10 "cL.z_rowmax_5dis_maj#cL.c_zmean_distance_5dis_cap"     "cL.z_rowmax_5dis_maj3#cL.c_zmean_distance_5dis_cap3"    "cL.z_rowmax_5dis_maj5#cL.c_zmean_distance_5dis_cap5"                       "cL.z_rowmax_5dis_maj10#cL.c_zmean_distance_5dis_cap10" L.lnpop  L.polity2) replace

esttab  D2model_1  D2model_3  D2model_5  D2model_10  D2model_11 D2model_13 D2model_15 D2model_20  using rich.rtf, compress nonotes r2 label eqlabels(none)  mtitles("yearly" "3 years" "5 years" "10 years" "yearly" "3 years" "5 years" "10 years" )  nogaps b(a2) se(a2) star(* 0.10 ** 0.05 *** 0.01) sfmt(%9.0fc) addnotes("Robust standard errors, clustered at the country level, are in parentheses. * p < 0.10, ** p < 0.05, *** p < 0.01.; country and year fixed effects are included; country-specific time trend is included") keep (L.z_rowmax_5dis_maj  L.z_rowmax_5dis_maj3   L.z_rowmax_5dis_maj5 L.z_rowmax_5dis_maj10 L.c_zmean_distance_5dis_cap L.c_zmean_distance_5dis_cap3 L.c_zmean_distance_5dis_cap5 L.c_zmean_distance_5dis_cap10 "cL.z_rowmax_5dis_maj#cL.c_zmean_distance_5dis_cap"     "cL.z_rowmax_5dis_maj3#cL.c_zmean_distance_5dis_cap3"    "cL.z_rowmax_5dis_maj5#cL.c_zmean_distance_5dis_cap5"                       "cL.z_rowmax_5dis_maj10#cL.c_zmean_distance_5dis_cap10" L.lnpop  L.polity2) replace


*********************Appendix 11. Estimating the effect of disasters on fiscal and administrative centralization, controlling political centralization

******

****11-1 correlation among three dimensions of decentralization*****



****estimation****
global controls2 "  l.lnpop  l.lngdppc l.polity2 l.political"


*average***
local i = 1
foreach depvar in fiscal administrative {
foreach indepvar in l.z_rowmax_5dis_maj l.z_rowmax_5dis_maj2 l.z_rowmax_5dis_maj3 l.z_rowmax_5dis_maj4 l.z_rowmax_5dis_maj5 l.z_rowmax_5dis_maj6 l.z_rowmax_5dis_maj7 l.z_rowmax_5dis_maj8 l.z_rowmax_5dis_maj9 l.z_rowmax_5dis_maj10{



reghdfe  `depvar' `indepvar' ${controls2} i.scode##c.year , absorb(scode year) cluster (scode)
est store polconmodel_`i'
est save polconmodel_`i', replace
local i = `i'+1

}
}



*****Figure 11-2. Coefficient plots of the effect of disasters  on institutional decentralization, control for RAI political decentralization*****
coefplot polconmodel_1 polconmodel_2 polconmodel_3 polconmodel_4 polconmodel_5 polconmodel_6 polconmodel_7 polconmodel_8 polconmodel_9 polconmodel_10 , keep (L.z_rowmax_5dis_maj L.z_rowmax_5dis_maj2 L.z_rowmax_5dis_maj3 L.z_rowmax_5dis_maj4 L.z_rowmax_5dis_maj5 L.z_rowmax_5dis_maj6 L.z_rowmax_5dis_maj7 L.z_rowmax_5dis_maj8 L.z_rowmax_5dis_maj9 L.z_rowmax_5dis_maj10) yline(0)  scheme(s2mono) graphregion(fcolor(white) ilcolor(white) lcolor(white)) vertical coeflabels (L.z_rowmax_5dis_maj="1" L.z_rowmax_5dis_maj2="2"  L.z_rowmax_5dis_maj3="3" L.z_rowmax_5dis_maj4="4" L.z_rowmax_5dis_maj5="5" L.z_rowmax_5dis_maj6="6" L.z_rowmax_5dis_maj7="7" L.z_rowmax_5dis_maj8="8" L.z_rowmax_5dis_maj9="9"  L.z_rowmax_5dis_maj10="10" ) title(fiscal) nokey
graph save polcontrolfiscal, replace




coefplot polconmodel_11 polconmodel_12 polconmodel_13 polconmodel_14 polconmodel_15 polconmodel_16 polconmodel_17 polconmodel_18 polconmodel_19 polconmodel_20  , keep (L.z_rowmax_5dis_maj L.z_rowmax_5dis_maj2 L.z_rowmax_5dis_maj3 L.z_rowmax_5dis_maj4 L.z_rowmax_5dis_maj5 L.z_rowmax_5dis_maj6 L.z_rowmax_5dis_maj7 L.z_rowmax_5dis_maj8 L.z_rowmax_5dis_maj9 L.z_rowmax_5dis_maj10) yline(0)  scheme(s2mono) graphregion(fcolor(white) ilcolor(white) lcolor(white)) vertical coeflabels (L.z_rowmax_5dis_maj="1" L.z_rowmax_5dis_maj2="2"  L.z_rowmax_5dis_maj3="3" L.z_rowmax_5dis_maj4="4" L.z_rowmax_5dis_maj5="5" L.z_rowmax_5dis_maj6="6" L.z_rowmax_5dis_maj7="7" L.z_rowmax_5dis_maj8="8" L.z_rowmax_5dis_maj9="9"  L.z_rowmax_5dis_maj10="10" ) title(administrative) nokey

graph save polcontroladmin, replace



graph combine polcontrolfiscal.gph polcontroladmin.gph  , row(1) ycommon scheme(s2mono) graphregion(fcolor(white) ilcolor(white) lcolor(white))

graph save polcontrolappend, replace 


*********Appendix 11-3. Regression coefficients table*******


esttab  polconmodel_1  polconmodel_3 polconmodel_5  polconmodel_10 polconmodel_11  polconmodel_13 polconmodel_15  polconmodel_20 using controlpolitical.rtf, compress nonotes r2 label eqlabels(none)  mtitles("yearly" "3 years" "5 years" "10 years" "yearly" "3 years" "5 years" "10 years" )  nogaps b(a2) se(a2) star(* 0.10 ** 0.05 *** 0.01) sfmt(%9.0fc) addnotes("Robust standard errors, clustered at the country level, are in parentheses. * p < 0.10, ** p < 0.05, *** p < 0.01.; country and year fixed effects are included; country-specific time trend is included") keep ( L.z_rowmax_5dis_maj  L.z_rowmax_5dis_maj3   L.z_rowmax_5dis_maj5 L.z_rowmax_5dis_maj10  L.lnpop  L.lngdppc L.polity2 L.political) replace


